Write Launch

Launch Copy — Orchestrator

3 agents (launch-copywriter → guard-checker → critic) emit a launch channel's complete native copy bundle from its launch-channel playbook pack, with the pack's §4 hard guards enforced. Capability: [routing.yaml](routing.yaml). Agents + dispatch + rubric: [references/agent-manifest.md](references/agent-manifest.md). Methodology: [references/playbook.md](references/playbook.md).

Core question: Does this bundle earn the launch — every native copy piece the channel needs, clearing every §2 cap and §4 hard guard, engineered to earn first-window velocity (never beg)? A launch is a bundle, not a post — the reason this is a separate emitter from write-social.

Critical Gates — load first

Full detail: [references/procedures/critical-gates.md](references/procedures/critical-gates.md).

1. Single-channel per artifact — multi-channel = re-invoke per channel (the launch chain fans out).
2. Pack-bound or transparent degrade — load the channel's launch-channel pack and narrate it (legibility). No pack → general principles only, stated as such; never fake channel tailoring.
3. Brand mode required (founder / company) — no silent default.
4. Hard guards are publish-blocking — the pack's §4 guards (PH: no vote-ask anywhere; Reddit: founder disclosure + 9:1 + sub-rule/flair fit) are non-negotiable. A breach unresolved in 1 guard-check revision = GUARD_FAIL → status: blocked, critic NOT dispatched.

Before Starting

Apply references/_shared/before-starting-check.md. Then:

| Artifact | Source | Required? | |---|---|---| | references/_shared/platform-intelligence/[channel].md | the pack | Hard for tailoring — else gate 2 degrade | | brand/BRAND.md | create-brand | Hard — brand_mode | | research/icp-research.md + product-context.md | research-icp | Rec — audience + voice + the differentiator | | docs/forsvn/artifacts/marketing/campaign-plan.md | plan-campaign | Opt — run-of-show + channel pick |

Mode per references/_shared/mode-resolver.md. --fast → guard loop = 0; --deep → MAX 2. --fast does NOT skip Cold Start, the hard guards, or the gates. Pre-Dispatch: [references/procedures/pre-dispatch.md](references/procedures/pre-dispatch.md).

Quality Gate — 5 dimensions

Full rubric + per-channel calibration + Discrimination Test: [references/rubric.md](references/rubric.md).

• [ ] Angle fit / clarity (identifier + anchor match a pack §1 angle)
• [ ] Format + hard-guard compliance (§2 caps AND §4 guards — a guard breach scores 0)
• [ ] Velocity-earning structure (earns first-window velocity per §3/§5; never a vote-ask)
• [ ] Channel-native voice / culture fit (reads native, not a press release)
• [ ] Bundle completeness + coherence (all components present + run-of-show consistent)

Pass total ≥35/50 AND no dim 0. Discrimination test every cycle.

Artifact Contract

• Path: docs/forsvn/artifacts/marketing/launch/[channel]-[YYYY-MM-DD]-[slug].md · Lifecycle: pipeline.
• Frontmatter (16): verbatim schema (adds legibility + guard_status) → [references/format-conventions.md](references/format-conventions.md).
• Body (in order): Launch bundle · Legibility block · Why this works · Critic verdict (6-row) · Anti-patterns triggered (- None if empty).
• Signature sidecars (FOR-46/U4): also emit .tagline.md + .first-comment.md (signature: ph-tagline/ph-first-comment), each reviewable alone — § Signature sidecars in format-conventions.

Side effects + consumed-by + cross-stack contract: [references/procedures/artifact-contract.md](references/procedures/artifact-contract.md).

Routing + Dispatch

Graph + fallback + polish + mode: [references/procedures/dispatch-mechanics.md](references/procedures/dispatch-mechanics.md). Guard-check bounce: PASSED → critic; REVISION_REQUIRED → copywriter ONCE; second → GUARD_FAIL (status: blocked, no critic).

Chain Position

Prev: plan-campaign (launch path) / research-icp / greenfield. Next: publish-social (distribute) → measure-results (post-launch read → pack write-back). The launch-copy step for a launch-channel pack — write-social owns it for the 5 social feed platforms. Re-run: new channel, voice shift, angle A/B, underperformance.

Deference: single social post (tweet / LinkedIn / TikTok caption) → write-social; run-of-show / channel pick / timing → plan-campaign; gallery / OG asset brief → brief-graphic; distributing finished copy → publish-social; post-launch results → measure-results; VN polish → polish-vn; AI-tell stripping → humanmaxxing.

Anti-Patterns

[references/anti-patterns.md](references/anti-patterns.md) — 10 patterns (6 launch-craft + 4 cross-cutting). Re-read before ship.

Durable Rules (protected)

Completion Status

• DONE — full bundle emitted, every §2 cap + §4 guard clear, critic ≥35, variant_count delivered.
• DONEWITHCONCERNS — critic 25-34 OR any dim <4 OR verdict fail (FAIL ships annotated) OR pack stale.
• BLOCKED — GUARD_FAIL (a §4 breach unresolved in 1 revision); brand_mode unresolved; no pack + operator declines the degrade.
• NEEDS_CONTEXT — no channel, topic, brand voice, or brand/BRAND.md; recommend plan-campaign or create-brand.

Execution

Registry-gated fork (category publish): Brief-only polish if flagged; Assisted/Direct need a verified engine. No auto-publish — review is human-owned. [execution-fork.md](references/_shared/execution-fork.md); record execution_mode.

Worked Example

Product Hunt founder-voice walkthrough (Cold Start → dispatch → no-vote-ask guard → critic PASS) + a Reddit value-first contrast: [references/examples/launch-walkthrough.md](references/examples/launch-walkthrough.md).

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: hungv47
• Source: hungv47/meta-skills
• License: MIT
• Homepage: https://skills.forsvn.com/

Install and usage instructions live in the source repository linked above.